- Email: [email protected]
Singular subsets of locus coeruleus neurons may recover tyrosine hydroxylase phenotype transiently expressed during development
Molecular Brain Research 76 Ž2000. 275–281 www.elsevier.comrlocaterbres
Research report
Singular subsets of locus coeruleus neurons may recover tyrosine hydroxylase phenotype transiently expressed during development Laurent Bezin ) , Dominique Marcel, Sebastien Desgeorges, Jean-Franc¸ois Pujol, ´ Dinah Weissmann Laboratoire de Neuropharmacologie Moleculaire, CNRSr UniÕersite´ Claude Bernard Lyon I r ERS 2022, Faculte´ de Medecine R.T.H. Laennec, ´ ´ ¨ Rue G. Paradin, F-69372 Lyon cedex 8, France Accepted 21 December 1999
Abstract The number of tyrosine hydroxylase ŽTH.-expressing neurons appears to be precisely determined in basal conditions within the noradrenergic pontine nucleus locus coeruleus ŽLC.. However, additional neurons exhibiting TH phenotype have been observed in the adult rat LC following a single administration of RU 24722, a potent inducer of TH expression specific to the LC. The neurons acquiring TH phenotype following treatment had a topographical localization similar to that of the neurons, which transiently expressed TH during postnatal development and lost TH phenotype during the third postnatal week. The idea that the fluctuation of TH phenotype in singular subsets of LC neurons during development may be selectively restored in adults is of particular interest. The present study attempted to determine whether the cells in which TH expression was repressed during the third postnatal week could correspond to those which exhibited TH phenotype in response to RU 24722 treatment in adults. We first verified that no massive cell death occurred in the LC during the period ranging from days 13 to 30. Then, we observed that both cell populations exhibited the same altered steady-state concentration of TH-mRNA as compared to cells that permanently expressed TH. Finally, we demonstrated the presence of TH-negative neurons expressing the homeodomain transcription factor Phox2a, specific for the determination of noradrenergic phenotype, providing further evidence that ‘‘resting-noradrenergic’’ neurons exist in the adult rat LC under basal conditions. These neurons provide interesting prospective for gain of noradrenergic function when classical noradrenergic LC neurons are impaired. q 2000 Elsevier Science B.V. All rights reserved. Keywords: Tyrosine hydroxylase; Phox2a; Noradrenergic phenotype; Development; Plasticity; Transcription factor; Locus coeruleus
1. Introduction Rat locus coeruleus ŽLC. neurons are densely packed within the pontine tegmentum and use noradrenaline in majority Žabout 90% of them. as their classical neurotransmitter w17x, the biosynthesis of which depends on the presence of the enzyme tyrosine hydroxylase ŽTH.. In basal conditions, the number of TH-expressing neurons and their spatial organization appear to be very constant from animal to animal. Surprisingly, additional neurons exhibited TH phenotype in the LC, 3 days after a single treatment with RU 24722. These additional TH-containing
) Corresponding author. Fax: q33-47-877-8732; e-mail: [email protected]
neurons were topographically organized at the immediate vicinity of neurons which constitutively expressed TH in basal conditions, particularly at the caudo-rostral junction of the nucleus ŽFig. 1A. w5x. The mechanism eliciting TH gene expression in response to RU 24722 remains unknown; however, this compound is now well characterized as an inducer of TH expression, specifically in the Žnor.adrenergic groups of the central nervous system Žreferences in Ref. w3x.. LC neurons appear to be spatially segregated within the nucleus with respect to the brain regions that they innervate. Neurons localized at the caudo-rostral junction of the LC target diverse brain regions such as spinal cord, cerebellum, hippocampus, hypothalamus and cortex w12x. Since the additional TH-expressing neurons revealed by RU 24722 treatment were mainly localized at this junction, it is likely that they target brain regions innervated by neigh-
0169-328Xr00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 3 2 8 X Ž 0 0 . 0 0 0 0 7 - 3
276
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
taining neurons, localized on coronal sections at the ventro-lateral tip of the rostral pole of the LC proper, lost TH phenotype between P14 and P21 ŽFig. 1B. w4x. These neurons may be those which acquire TH expression in adult rats following RU 24722 treatment w5x, since Ž1. they exhibit a very similar spatial organization within the nucleus ŽFig. 1A,B. and Ž2. the number of additional THcontaining neurons in adult rats treated with RU 24722 was similar to the excess number of TH-expressing neurons at P14 Ž; 30% as compared to adult control rats.. Considering the importance that such a plasticity of TH phenotype may have in the functions involving LC neurons, the present study was devoted to seeking further evidence showing that cells, which transiently exhibited TH phenotype during postnatal development, may constitute in adults singular subsets of LC neurons, in which TH phenotype may be triggered by specific stimuli.
2. Materials and methods 2.1. Animals
Fig. 1. Plasticity of TH phenotype in adult ŽA. and young ŽB. rat LC. ŽA. Sagittal representation of the LC in adult rats ŽP42., illustrating the location of the additional cells Žopen-stars., which exhibited TH phenotype in response to RU 24722 treatment w5x. At the caudo-rostral junction of the nucleus, demarcated by the arrow, these cells caused an ‘‘enlargement’’ of the area occupied by TH-expressing cells within the nucleus, as quantitatively demonstrated on coronal sections Žreproduced from Ref. w5x.. Cells expressing constitutively TH in basal conditions Žfilled circles. delineate the ‘‘core area’’ of the LC. 4v: 4th ventricle. ŽB. Two populations of LC neurons could be distinguished at P14 based on the time when they lost TH phenotype during late postnatal development and their localization w4x. One population Žopen squares. Ža. lost TH phenotype between P21 and P42 and Žb. was localized in the ‘‘core area of the LC’’, the limits of which are defined by the presence of the cells that permanently expressed TH throughout development Žv .. The other population of cells Ždark stars., Ža. lost TH phenotype between P14 and P21, Žb. was localized at the vicinity of the ‘‘core area of the LC’’, defining the so-called ‘‘extra-core area of the LC’’, and Žc. was reminiscent of the cells Žopen-stars, A., which acquired TH phenotype in response to RU 24722 treatment.
boring LC neurons. Therefore, the recovery of noradrenergic function in these neurons may be useful to compensate for impaired noradrenergic neurons in adults. Such a plasticity of TH phenotype elicited in adult rats is reminiscent of the fluctuating expression of TH phenotype that occurs in the LC during postnatal development. In 14-day-old ŽP14. rats, a singular population of TH-con-
OFA male rats ŽIffa-Credo, Lyon, France. were kept 1 week under 12:12 h darkrlight cycle at 218C with food and water ad libitum. Rats treated with RU 24722 Žobtained from Dr. C. Thal. received an i.p. injection of a 30 mgrkg solution, pH 2.8, prepared in 0.04N HCl. Control rats of the same age received an i.p. injection of 0.004 N HCl, pH 2.8. Rats were killed 24 or 72 h after the injection. 2.2. In situ end-labeling (ISEL) of DNA Cryostat-cut coronal sections Ž20 mm. of fresh-frozen brains Žrats’ ages: P13, P14, P16, P19, P21, P23 and P30. were selected in the most rostral plane of LC proper Žanatomical plane 9 w4x. by reference to adjacent Nisslstained sections, collected on gelatin-coated slides, postfixed 20 min at 48C in a solution of 4% paraformaldehyde ŽPF. in 0.1 M phosphate buffer saline, pH 7.4 ŽPBS., incubated overnight at 48C in PBS containing 1% bovine serum albumin ŽBSA. and 0.3% Triton X-100, and then processed using Klenow fragment of DNA polymerase to incorporate biotinylated nucleotides into dsDNA breaks formed during apoptosis, and incubated in streptavidinconjugated horseradish peroxidase ŽHRP., following the manufacturer’s instructions ŽTacs, Trevigen.. HRP was detected using diaminobenzidine ŽDAB, Sigma.. Sections were then Nissl-stained with Cresyl violet. 2.3. Immunohistochemistry Brains of rats ŽP14 and P42. transcardially perfused with an ice-cold solution of 4% PF in 0.1 M PBS, cryopro-
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
tected in PBS with 25% sucrose at 48C, were frozen in isopentane. Coronal cryostat-cut sections Ž20 mm. were selected throughout the caudo-rostral extent of the LC. TH and Phox2a detections were performed using a mouse monoclonal anti-TH antibody ŽIncstar, 1:3000 dilution. and a rabbit anti-Phox2a antibody Žgifted by Dr. C. Goridis, 1:1000 dilution., biotinylated antibodies raised against mouse or rabbit IgG ŽVector., respectively, and peroxydase-vectastain Elite ABC reagents ŽVector. and DAB. An image analysis system ŽImaging Res.. allowed Ž1. to determine the surface area Žmm2 . demarcated by TH-immunopositive ŽTH-IP. perikarya, Ž2. to measure the diameter of Phox2a-immunopositive ŽPhox2a-IP. nuclei and to count these nuclei at 200-mm intervals throughout the caudo-rostral axis of the LC. The total number of Phox2a-IP cells was estimated in the LC using the Abercrombie formula w1x. 2.4. Topological reconstructions Co-alignment of sections selected at the rostral pole of the LC and belonging to different brains was performed using different landmarks surrounding the LC. The mean limits" S.E.M. of the structures considered were reconstructed in a reference system, as already described in details w4x. 2.5. RadioactiÕe in situ hybridization of TH-mRNA Cryostat-cut coronal sections Ž10 mm. of fresh-frozen brains Žrats’ ages: P14, P21 and P23. were selected in the most rostral plane of LC proper Žanatomical plane 9 w4x. by reference to adjacent Nissl-stained sections, collected on Superfrost Plus slides ŽMenzel-Glaser ¨ .. In this study, sections were processed all together. In situ hybridization was performed as previously described w3,6x, using a synthetic oligonucleotidic probe complementary to the coding sequence Žnucleotide 1233 to 1267. of rat TH w8x, radiolabeled by ‘‘tailing’’ the 3X end with w a y35 Sx deoxyadenosine triphosphate. Specific activity of the labeled oligomer was 8 = 10 6 c.p.m.rpmol. 2.5.1. Determination of TH-mRNA tissue concentration Film autoradiograms, obtained by exposing slides to b-max hyperfilms, were analyzed by densitometry, providing an index of TH-mRNA tissue concentration. Statistical analysis was performed using two-way analysis of variance, followed by Newman–Keuls post-hoc test. 2.5.2. Determination of TH-mRNA perikaryal concentration Using an image analysis system ŽImaging Res.., specific silver grain density over LC neurons was determined as previously described w3x, following development of
277
emulsion-coated slides and counterstaining with Cresyl violet. This density represents an index of the perikaryal concentration of TH-mRNA, expressed in arbitrary units ŽAU.. One AU corresponds to 1% of the perikarya surface occupied by silver grains. Since slices were processed altogether, it was possible to compare all the different measures. 2.6. Combined non-radioactiÕe in situ hybridization and immunohistochemistry Using a Boehringer transcription kit, T7 polymerase was used to transcribe a digoxygenin ŽDIG.-labeled antisense cRNA probe from a ; 1.5 kb KpnI–BamHI fragment of plasmid pSPT-RTH Žgifted by Dr. J. Mallet. that is complementary to part of the coding region of rat TH-mRNA w8x. P42 rat brains were prepared as described in Section 2.3. Coronal sections Ž20 mm thick. were collected on Superfrost Plus slides, rinsed in PBS, incubated for 15 min in PBS containing 0.3% Triton X-100, thrice rinsed in PBS, incubated for 15 min in 5 = SSC at 378C, and hybridized overnight at 708C in hybridization buffer Ž50% formamide, 5 = SSC, 5 = Denhardt’s, 250 mgrml yeast tRNA, 500 mgrml salmon sperm DNA. containing 0.5 ngrml heat-denaturated DIG-labeled cRNA probe. Slides were then processed as previously described w18x. DIG was detected using alkaline phosphatase-conjugated anti-DIG antibody, 4-nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indoyl-phosphate. Sections were then processed for Phox2a immunohistochemistry, as described above.
3. Results 3.1. Apoptosis in the LC during postnatal deÕelopment Seven postnatal ages ŽP13, P14, P16, P19, P21, P23, and P30. were selected to monitor, in the rostral pole of the LC proper, the presence of cell death, by labeling DNA breaks using the ISEL method. Among all sections analyzed, three cells only exhibiting DNA breaks were observed in the LC, one at P14 ŽFig. 2A. and one at P16 in the ‘‘core area of the LC’’, and one at P16 in the ‘‘extracore area’’ of the LC. The ponctuata appearance of these dying cells and the high condensation of the labeled material ŽFig. 2B. indicate that cell death likely occurred via apoptosis. 3.2. TH-mRNA concentration in subsets of neurons in the rostral pole of LC proper at P14 The detection of TH-mRNA by in situ hybridization at P23 ŽFig. 3B. confirmed the disappearance of TH gene
278
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
tissue concentration increased 66 " 23% Ž p - 0.05. and 110 " 28% Ž p - 0.01. above controls, 1 and 3 days after treatment, respectively. Three days after treatment, at the rostral pole of the LC proper, the number of TH-mRNAcontaining perikarya increased, causing an ‘‘enlargement’’ of the nucleus and the emergence of an ‘‘extra-core area’’ of the LC ŽFig. 3B,C.. The density of specific silver grains in the LC perikarya, which was used as an index of TH-mRNA perikaryal concentration, was normally distributed across the populations of cells selected in the ‘‘core area of the LC’’, in control rats Žnot shown. and RU 24722-treated rats Ž n s 52, Pearson x 2 test of normality, Fig. 3F.. In the ‘‘extra-core area’’ of the LC, which was revealed in RU 24722-treated rats, the silver grain density was not normally distributed ŽFig. 3G.. Fig. 2. In situ detection of apoptotic neurons in the rostral pole of the LC at P14. ŽA. On this Nissl-stained section, a single apoptotic neuron Žpointed by the arrow. was detected in the LC Žoutlined area., and was located in the ‘‘core area’’ of the nucleus. ŽB. Observation of the apoptotic neuron pointed in panel A at a higher magnification, demonstrating dense condensation of the DNA. Bars: 100 mm ŽA.; 20 mm ŽB..
expression in the cells which are localized, at P14, in the ventro-lateral tip of the LC, the so-called ‘‘extra-core area’’ ŽFig. 3A.. The density of silver grains corresponding to non-specific labeling was determined in areas ventral to the LC and ranged between 0.8% and 1.4% of the silver grain density determined in the perikarya of LC neurons. The density of specific silver grains in the perikarya selected randomly for evaluation was normally distributed across the population of cells selected in the ‘‘core area of the LC’’ Ž n s 52, Pearson x 2 test of normality. ŽFig. 3D., but this density was not normally distributed in the ‘‘extra-core area’’ of the LC ŽFig. 3E.. 3.3. TH gene expression in response to RU 24722 treatment at P14 and P23 We investigated in P14 RU 24722-treated rats, whether additional TH-expressing cells could be detected at the vicinity of LC cells which expressed TH gene in basal conditions. The topology of the LC delineated by TH-IP perikarya was reconstructed at P14 in control rats and in rats which had received a single dose of RU 24722 at P11. We found that no alteration had occurred in RU 24722treated rats, both in the surface area delineated by TH-IP perikarya and the topology of the LC as compared to controls Žnot shown.. Pharmacologically induced alterations in TH-mRNA tissue concentration in the LC were monitored at P23 by in situ hybridization and film autoradiography. TH-mRNA tissue concentration was measured 1 and 3 days after a single administration of RU 24722, in four anatomical planes selected at the caudo-rostral junction of the LC and separated by 160 mm. In this 640-mm interval, TH-mRNA
3.4. Phox2a expression in the LC of control and RU 24722-treated adult rats The anatomical distribution of Phox2a-IP cells was heterogeneous along the caudo-rostral axis of the LC; the majority Ž; 80%. of stained cells was contained in the half-caudal subdivision of the structure Žanatomical planes 1 to 4, Fig. 4A.. The mean diameter d of labeled nuclei was 10.7 mm Ž n s 71; 8.2 - d - 13.1 mm.. The number of Phox2a-IP cells, estimated in the entire structure using the Abercrombie formula, was 2588 " 164 in basal conditions. In situ dual labeling of TH-mRNA and Phox2a protein revealed that all TH-IP neurons expressed Phox2a and that numerous Phox2a-IP cells did not contain detectable level of TH-mRNA; 47% of total Phox2a-IP cells did not contain TH at anatomical plane 4 ŽFig. 4B.. RU 24722 treatment did not change the caudo-rostral distribution of Phox2a-IP cells in the LC ŽFig. 4A., nor the total number of Phox2a-IP cells contained in the entire structure Ž2453 " 122..
4. Discussion 4.1. Apoptotic death is not preÕalent in the deÕeloping LC Excess numbers of neurons are produced during the ontogenesis within the central nervous system, and are eliminated, often having served important but transient functions w15x. The noradrenergic innervation of the brain is progressive during pre- and postnatal development, and the noradrenaline released by LC neurons may be involved in the critical periods of brain plasticity and regulate different brain development steps w2,10,13,16x. We previously reported that the number of TH-expressing neurons at P14 was 33% greater than that at P42 w4x. If the function served by these additional TH-expressing neurons was transient during development, the loss of TH phenotype in
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
279
Fig. 3. Analysis of TH-mRNA presence at the rostral pole of LC proper. ŽA–C. Polarized light views of emulsion-coated coronal sections processed for TH-mRNA in situ hybridization using a 35 S-labelled oligonucleotidic probe, at P14 ŽA. and at P23 ŽB, C., in control ŽA, B. and RU 24722-treated rats ŽC.. Dashed lines indicate the position of the ventral limit of the 4th ventricle. Bar: 300 mm ŽA, B, C.. ŽD–G. Frequency histograms illustrating the distribution of an index of TH-mRNA perikaryal concentration from individual perikarya selected within the ‘‘core area’’ ŽD, F. and the ‘‘extra-core area’’ ŽE, G., in P14 control rats ŽD, E. and P23 RU 24722-treated rats ŽF, G.. At P23, the ‘‘core area’’ only was detected in control rats and its topology has been quantitatively reconstructed. The superimposition of this reconstructed ‘‘core area’’ over each section analyzed at P14, and at P23 in RU 24722-treated rats, made it possible to isolate the ‘‘core area’’ and the ‘‘extra-core area’’ of the LC. Index of TH-mRNA perikaryal concentration was expressed as the density of silver grains ŽAU. over selected neuron perikarya of four animals. Total number n of neurons considered is indicated. The index of TH-mRNA perikaryal concentration was ‘‘normally’’ distributed across the population of neurons selected in the ‘‘core area’’ of the LC only.
these neurons beyond P14 may reflect either TH gene repression or massive cell death. Three apoptotic cells only were detected at the rostral pole of the LC proper throughout the time period of postnatal development examined ŽP13 to P30.. Since numerous apoptotic cells were observed at P13 in the substantia nigra Žnot shown., as previously reported w9x, the methodology we used can be considered as sensitive and reliable enough to detect apoptotic cells in brain sections. No cells were double-labeled
when dual ISEL and immunohistochemical detection of TH was performed. This may indicate that dying cells did not express TH or that the epitopes of TH protein recognized by the different antibodies raised against TH that we used were already altered when immunohistochemistry was performed. Anyway, the present data strongly suggest that apoptotic cell death was not a massive phenomenon within the rostral pole of the LC proper throughout the postnatal development of the rat. Therefore, loss of TH
280
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
phenotype beyond P14 may reflect TH gene repression rather than cell death. 4.2. TH-mRNA concentration reÕeals transient TH phenotype in subsets of LC neurons We investigated whether it was possible to predict transient expression of TH phenotype in the ‘‘extra-core area’’ of the LC in development or in adults treated with RU 24722. Therefore, we compared some modalities of TH gene expression in the ‘‘extra-core area’’ with that of the ‘‘core area’’, which contains LC neurons expressing TH permanently in basal conditions throughout development. TH-mRNA was chosen as a marker to examine, since, in contrast with TH protein, its subcellular distribution is restricted to the perikarya of rat LC neurons w3x, making it possible to measure its concentration in individual neurons. We observed that the concentration of THmRNA exhibited a profile of ‘‘normal’’ distribution in the population of cells localized in the ‘‘core area’’ of the LC, in any conditions examined, but not in the ‘‘extra-core area’’ of the LC both at P14 in control LC and at P23 in the LC of rats treated with RU 24722 at P20. The observation of a bi-modal profile in the ‘‘extra-core area’’ probably means that neurons were progressively recruited for transcription of the TH gene, as supported by the observation that there is no synchronous transcription activity of TH gene among the population of LC neurons w19x. In the ‘‘extra-core area’’ of the LC, cells with low concentration of TH-mRNA may have just started to express TH or may have already started the process of TH gene repression at the time when measurements were performed. 4.3. Expression of homeoprotein Phox2a in ‘‘restingnoradrenergic’’ adult LC neurons
Fig. 4. Expression of Phox2a homeoprotein in the LC of adult rats in basal conditions and 3 days after a single treatment with RU 24722. ŽA. Number of Phox2a-IP cells counted in 20-mm-thick sections, selected every 200 mm along the caudo-rostral extent of the LC. Results are expressed as the means"S.E.M. Ž ns 3., for both control Žwhite bars. and RU 24722-treated Ždark bars. rats. ŽB. Color photomicrograph of a section through the rostral LC of control rat, from material processed for dual in situ hybridization of TH-mRNA and immunohistochemical detection of Phox2a. TH-mRNAq cells are apparent by the presence of blue reaction product in their cytoplasm and Phox2aq cells by that of brown reaction product in their nucleus. At this anatomical plane Žnumber 4, as in A., Phox2a-IP cells without detectable expression of TH-mRNA are pointed by arrowheads. 4v: 4th ventricle. Bar: 50 mm.
Homeodomain transcription factor Phox2a is expressed in domains within the nervous system corresponding to noradrenergic cell populations w18x, and appears to be a determinant of the noradrenergic phenotype w7,14x. We found in the present study that numerous cells in the LC expressed Phox2a, but not TH in basal conditions, suggesting the existence of a ‘‘reservoir’’ of cells within the LC with the intrinsic capacity to develop noradrenergic phenotype in response to pertinent stimuli. Indeed, we previously reported ; 1700 TH-positive neurons in the adult rat LC w4,5x, contrasting with the ; 2600 Phox2a-positive cells counted in the present study. The observation that the number of Phox2a-expressing cells did not increase following RU 24722 treatment, as does the number of TH-expressing cells w5x, indicates that TH gene expression may have been triggered in response to the treatment in neurons that already expressed Phox2a in basal conditions. Interestingly, the number of Phox2a positive cells was greater than the number Ž; 2200 w5x. of TH-positive cells counted in the LC following RU 24722. Thus, subsets of Phox2a-ex-
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
pressing cells still do not express TH in response to RU 24722 treatment, and may correspond to the cells that lost TH phenotype between P21 and P42 ŽFig. 1B.. Thus, Phox2a is not sufficient to activate TH gene expression in singular subsets of LC neurons in basal conditions. However, Phox2a has been demonstrated to induce expression of TH in collaboration with cAMP w11x. Further studies will be needed to evaluate the involvement of cAMP in the plasticity of TH phenotype in LC neurons expressing Phox2a. 5. Conclusion In attempting to determine whether the cells in which TH expression was repressed between P14 and P21 could correspond to those which exhibited TH phenotype in response to RU 24722 treatment in adults, we verified that no massive cell death occurred in the LC during the postnatal period. We also observed that both cell populations exhibited the same altered steady-state concentration of TH-mRNA as compared to cells which permanently expressed TH. Finally, we have demonstrated the presence of TH-negative neurons expressing the homeodomain transcription factor Phox2a, specific for noradrenergic neurons, providing further evidence that ‘‘restingnoradrenergic’’ neurons exist in the adult rat LC under basal conditions. The existence of these numerous cells, localized in the LC where efferents have been shown to be widely spread across the central nervous system, provides interesting prospectives for gain of noradrenergic function when classical noradrenergic LC neurons are impaired. Acknowledgements This work was supported by grants from the Centre National de la Recherche Scientifique ŽERS 2022. and the Universite´ Claude Bernard Lyon I. The authors are very grateful to Drs. C. Goridis and J.-Fr. Brunet for their critical reading of the manuscript. References w1x M. Abercrombie, Estimation of nuclear population from microtome sections, Anat. Rec. 94 Ž1946. 239–247. w2x M.F. Bear, W. Singer, Modulation of visual cortical plasticity by acetylcholine and noradrenaline, Nature 320 Ž1986. 172–176.
281
w3x L. Bezin, J.J. Diaz, D. Marcel, M. Le Cavorsin, J.J. Madjar, J.F. Pujol, D. Weissmann, Controlled targeting of tyrosine hydroxylase protein toward processes of locus coeruleus neurons during postnatal development, Mol. Brain Res. 50 Ž1997. 23–32. w4x L. Bezin, D. Marcel, L.I. Debure, N. Ginovart, C. Rousset, J.F. Pujol, D. Weissmann, Postnatal development of the tyrosine hydroxylase-containing cell population within the rat locus coeruleus: topological organization and phenotypic plasticity, J. Neurosci. 14 Ž1994. 7486–7501. w5x L.I. Debure, L. Bezin, N. Ginovart, C. Rousset, J.F. Pujol, D. Weissmann, RU24722 induces spatially organized phenotypic plasticity in the locus coeruleus, NeuroReport 5 Ž1994. 1793–1796. w6x L.I. Debure, E. Moyse, M. Fevre-Montange, H. Hardin, M.F. Belin, C. Rousset, J.F. Pujol, D. Weissmann, Somatotopic organization of tyrosine hydroxylase in the rat locus coeruleus: long-term effect of RU24722, Brain Res. 581 Ž1992. 19–32. w7x C. Goridis, J.F. Brunet, Transcriptional control of neurotransmitter phenotype, Curr. Opin. Neurobiol. 9 Ž1999. 47–53. w8x B. Grima, A. Lamouroux, F. Blanot, N. Faucon-Biguet, J. Mallet, Complete coding sequence of rat tyrosine hydroxylase mRNA, Proc. Natl. Acad. Sci. U. S. A. 82 Ž1985. 617–621. w9x E. Janec, R.E. Burke, Naturally occurring cell death during postnatal development of the substantia nigra pars compacta of rat, Mol. Cell. Neurosci. 4 Ž1993. 30–35. w10x T. Kasamatsu, J.D. Pettigrew, Preservation of binocularity after monocular deprivation in the striate cortex of kittens treated with 6-hydroxydopamine, J. Comp. Neurol. 185 Ž1979. 139–162. w11x L. Lo, X. Morin, J.F. Brunet, D.J. Anderson, Specification of neurotransmitter identity by Phox2 proteins in neural crest stem cells, Neuron 22 Ž1999. 693–705. w12x S.E. Loughlin, S.L. Foote, F.E. Bloom, Efferent projections of nucleus locus coeruleus: topographic organization of cells of origin demonstrated by three-dimensional reconstruction, Neuroscience 18 Ž1986. 291–306. w13x T. Maeda, M. Tohyama, N. Shimizu, Modification of postnatal development of neocortex in rat brain with experimental deprivation of locus coeruleus, Brain Res. 70 Ž1974. 515–520. w14x X. Morin, H. Cremer, M.R. Hirsch, R.P. Kapur, C. Goridis, J.F. Brunet, Defects in sensory and autonomic ganglia and absence of locus coeruleus in mice deficient for the homeobox gene Phox2a, Neuron 18 Ž1997. 411–423. w15x R.W. Oppenheim, Cell death during development of the nervous system, Annu. Rev. Neurosci. 14 Ž1991. 453–501. w16x J.G. Parnavelas, M.E. Blue, The role of the noradrenergic system on the formation of synapses in the visual cortex of the rat, Dev. Brain Res. 3 Ž1982. 140–144. w17x L.W. Swanson, The locus coeruleus: a cytoarchitectonic, golgi and immunocytochemical study in the albino rat, Brain Res. 110 Ž1976. 39–56. w18x M.C. Tiveron, M.R. Hirsch, J.F. Brunet, The expression pattern of the transcription factor Phox2 delineates synaptic pathways of the autonomic nervous system, J. Neurosci. 16 Ž1996. 7649–7660. w19x A. Trembleau, F.E. Bloom, Subcellular localization of tyrosine hydroxylase ŽTH. gene transcripts: new insights into the pattern of TH gene expression in the locus coeruleus under pharmacological stimulation, Biol. Cell 90 Ž1998. 39–51.
Research report
Singular subsets of locus coeruleus neurons may recover tyrosine hydroxylase phenotype transiently expressed during development Laurent Bezin ) , Dominique Marcel, Sebastien Desgeorges, Jean-Franc¸ois Pujol, ´ Dinah Weissmann Laboratoire de Neuropharmacologie Moleculaire, CNRSr UniÕersite´ Claude Bernard Lyon I r ERS 2022, Faculte´ de Medecine R.T.H. Laennec, ´ ´ ¨ Rue G. Paradin, F-69372 Lyon cedex 8, France Accepted 21 December 1999
Abstract The number of tyrosine hydroxylase ŽTH.-expressing neurons appears to be precisely determined in basal conditions within the noradrenergic pontine nucleus locus coeruleus ŽLC.. However, additional neurons exhibiting TH phenotype have been observed in the adult rat LC following a single administration of RU 24722, a potent inducer of TH expression specific to the LC. The neurons acquiring TH phenotype following treatment had a topographical localization similar to that of the neurons, which transiently expressed TH during postnatal development and lost TH phenotype during the third postnatal week. The idea that the fluctuation of TH phenotype in singular subsets of LC neurons during development may be selectively restored in adults is of particular interest. The present study attempted to determine whether the cells in which TH expression was repressed during the third postnatal week could correspond to those which exhibited TH phenotype in response to RU 24722 treatment in adults. We first verified that no massive cell death occurred in the LC during the period ranging from days 13 to 30. Then, we observed that both cell populations exhibited the same altered steady-state concentration of TH-mRNA as compared to cells that permanently expressed TH. Finally, we demonstrated the presence of TH-negative neurons expressing the homeodomain transcription factor Phox2a, specific for the determination of noradrenergic phenotype, providing further evidence that ‘‘resting-noradrenergic’’ neurons exist in the adult rat LC under basal conditions. These neurons provide interesting prospective for gain of noradrenergic function when classical noradrenergic LC neurons are impaired. q 2000 Elsevier Science B.V. All rights reserved. Keywords: Tyrosine hydroxylase; Phox2a; Noradrenergic phenotype; Development; Plasticity; Transcription factor; Locus coeruleus
1. Introduction Rat locus coeruleus ŽLC. neurons are densely packed within the pontine tegmentum and use noradrenaline in majority Žabout 90% of them. as their classical neurotransmitter w17x, the biosynthesis of which depends on the presence of the enzyme tyrosine hydroxylase ŽTH.. In basal conditions, the number of TH-expressing neurons and their spatial organization appear to be very constant from animal to animal. Surprisingly, additional neurons exhibited TH phenotype in the LC, 3 days after a single treatment with RU 24722. These additional TH-containing
) Corresponding author. Fax: q33-47-877-8732; e-mail: [email protected]
neurons were topographically organized at the immediate vicinity of neurons which constitutively expressed TH in basal conditions, particularly at the caudo-rostral junction of the nucleus ŽFig. 1A. w5x. The mechanism eliciting TH gene expression in response to RU 24722 remains unknown; however, this compound is now well characterized as an inducer of TH expression, specifically in the Žnor.adrenergic groups of the central nervous system Žreferences in Ref. w3x.. LC neurons appear to be spatially segregated within the nucleus with respect to the brain regions that they innervate. Neurons localized at the caudo-rostral junction of the LC target diverse brain regions such as spinal cord, cerebellum, hippocampus, hypothalamus and cortex w12x. Since the additional TH-expressing neurons revealed by RU 24722 treatment were mainly localized at this junction, it is likely that they target brain regions innervated by neigh-
0169-328Xr00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 3 2 8 X Ž 0 0 . 0 0 0 0 7 - 3
276
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
taining neurons, localized on coronal sections at the ventro-lateral tip of the rostral pole of the LC proper, lost TH phenotype between P14 and P21 ŽFig. 1B. w4x. These neurons may be those which acquire TH expression in adult rats following RU 24722 treatment w5x, since Ž1. they exhibit a very similar spatial organization within the nucleus ŽFig. 1A,B. and Ž2. the number of additional THcontaining neurons in adult rats treated with RU 24722 was similar to the excess number of TH-expressing neurons at P14 Ž; 30% as compared to adult control rats.. Considering the importance that such a plasticity of TH phenotype may have in the functions involving LC neurons, the present study was devoted to seeking further evidence showing that cells, which transiently exhibited TH phenotype during postnatal development, may constitute in adults singular subsets of LC neurons, in which TH phenotype may be triggered by specific stimuli.
2. Materials and methods 2.1. Animals
Fig. 1. Plasticity of TH phenotype in adult ŽA. and young ŽB. rat LC. ŽA. Sagittal representation of the LC in adult rats ŽP42., illustrating the location of the additional cells Žopen-stars., which exhibited TH phenotype in response to RU 24722 treatment w5x. At the caudo-rostral junction of the nucleus, demarcated by the arrow, these cells caused an ‘‘enlargement’’ of the area occupied by TH-expressing cells within the nucleus, as quantitatively demonstrated on coronal sections Žreproduced from Ref. w5x.. Cells expressing constitutively TH in basal conditions Žfilled circles. delineate the ‘‘core area’’ of the LC. 4v: 4th ventricle. ŽB. Two populations of LC neurons could be distinguished at P14 based on the time when they lost TH phenotype during late postnatal development and their localization w4x. One population Žopen squares. Ža. lost TH phenotype between P21 and P42 and Žb. was localized in the ‘‘core area of the LC’’, the limits of which are defined by the presence of the cells that permanently expressed TH throughout development Žv .. The other population of cells Ždark stars., Ža. lost TH phenotype between P14 and P21, Žb. was localized at the vicinity of the ‘‘core area of the LC’’, defining the so-called ‘‘extra-core area of the LC’’, and Žc. was reminiscent of the cells Žopen-stars, A., which acquired TH phenotype in response to RU 24722 treatment.
boring LC neurons. Therefore, the recovery of noradrenergic function in these neurons may be useful to compensate for impaired noradrenergic neurons in adults. Such a plasticity of TH phenotype elicited in adult rats is reminiscent of the fluctuating expression of TH phenotype that occurs in the LC during postnatal development. In 14-day-old ŽP14. rats, a singular population of TH-con-
OFA male rats ŽIffa-Credo, Lyon, France. were kept 1 week under 12:12 h darkrlight cycle at 218C with food and water ad libitum. Rats treated with RU 24722 Žobtained from Dr. C. Thal. received an i.p. injection of a 30 mgrkg solution, pH 2.8, prepared in 0.04N HCl. Control rats of the same age received an i.p. injection of 0.004 N HCl, pH 2.8. Rats were killed 24 or 72 h after the injection. 2.2. In situ end-labeling (ISEL) of DNA Cryostat-cut coronal sections Ž20 mm. of fresh-frozen brains Žrats’ ages: P13, P14, P16, P19, P21, P23 and P30. were selected in the most rostral plane of LC proper Žanatomical plane 9 w4x. by reference to adjacent Nisslstained sections, collected on gelatin-coated slides, postfixed 20 min at 48C in a solution of 4% paraformaldehyde ŽPF. in 0.1 M phosphate buffer saline, pH 7.4 ŽPBS., incubated overnight at 48C in PBS containing 1% bovine serum albumin ŽBSA. and 0.3% Triton X-100, and then processed using Klenow fragment of DNA polymerase to incorporate biotinylated nucleotides into dsDNA breaks formed during apoptosis, and incubated in streptavidinconjugated horseradish peroxidase ŽHRP., following the manufacturer’s instructions ŽTacs, Trevigen.. HRP was detected using diaminobenzidine ŽDAB, Sigma.. Sections were then Nissl-stained with Cresyl violet. 2.3. Immunohistochemistry Brains of rats ŽP14 and P42. transcardially perfused with an ice-cold solution of 4% PF in 0.1 M PBS, cryopro-
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
tected in PBS with 25% sucrose at 48C, were frozen in isopentane. Coronal cryostat-cut sections Ž20 mm. were selected throughout the caudo-rostral extent of the LC. TH and Phox2a detections were performed using a mouse monoclonal anti-TH antibody ŽIncstar, 1:3000 dilution. and a rabbit anti-Phox2a antibody Žgifted by Dr. C. Goridis, 1:1000 dilution., biotinylated antibodies raised against mouse or rabbit IgG ŽVector., respectively, and peroxydase-vectastain Elite ABC reagents ŽVector. and DAB. An image analysis system ŽImaging Res.. allowed Ž1. to determine the surface area Žmm2 . demarcated by TH-immunopositive ŽTH-IP. perikarya, Ž2. to measure the diameter of Phox2a-immunopositive ŽPhox2a-IP. nuclei and to count these nuclei at 200-mm intervals throughout the caudo-rostral axis of the LC. The total number of Phox2a-IP cells was estimated in the LC using the Abercrombie formula w1x. 2.4. Topological reconstructions Co-alignment of sections selected at the rostral pole of the LC and belonging to different brains was performed using different landmarks surrounding the LC. The mean limits" S.E.M. of the structures considered were reconstructed in a reference system, as already described in details w4x. 2.5. RadioactiÕe in situ hybridization of TH-mRNA Cryostat-cut coronal sections Ž10 mm. of fresh-frozen brains Žrats’ ages: P14, P21 and P23. were selected in the most rostral plane of LC proper Žanatomical plane 9 w4x. by reference to adjacent Nissl-stained sections, collected on Superfrost Plus slides ŽMenzel-Glaser ¨ .. In this study, sections were processed all together. In situ hybridization was performed as previously described w3,6x, using a synthetic oligonucleotidic probe complementary to the coding sequence Žnucleotide 1233 to 1267. of rat TH w8x, radiolabeled by ‘‘tailing’’ the 3X end with w a y35 Sx deoxyadenosine triphosphate. Specific activity of the labeled oligomer was 8 = 10 6 c.p.m.rpmol. 2.5.1. Determination of TH-mRNA tissue concentration Film autoradiograms, obtained by exposing slides to b-max hyperfilms, were analyzed by densitometry, providing an index of TH-mRNA tissue concentration. Statistical analysis was performed using two-way analysis of variance, followed by Newman–Keuls post-hoc test. 2.5.2. Determination of TH-mRNA perikaryal concentration Using an image analysis system ŽImaging Res.., specific silver grain density over LC neurons was determined as previously described w3x, following development of
277
emulsion-coated slides and counterstaining with Cresyl violet. This density represents an index of the perikaryal concentration of TH-mRNA, expressed in arbitrary units ŽAU.. One AU corresponds to 1% of the perikarya surface occupied by silver grains. Since slices were processed altogether, it was possible to compare all the different measures. 2.6. Combined non-radioactiÕe in situ hybridization and immunohistochemistry Using a Boehringer transcription kit, T7 polymerase was used to transcribe a digoxygenin ŽDIG.-labeled antisense cRNA probe from a ; 1.5 kb KpnI–BamHI fragment of plasmid pSPT-RTH Žgifted by Dr. J. Mallet. that is complementary to part of the coding region of rat TH-mRNA w8x. P42 rat brains were prepared as described in Section 2.3. Coronal sections Ž20 mm thick. were collected on Superfrost Plus slides, rinsed in PBS, incubated for 15 min in PBS containing 0.3% Triton X-100, thrice rinsed in PBS, incubated for 15 min in 5 = SSC at 378C, and hybridized overnight at 708C in hybridization buffer Ž50% formamide, 5 = SSC, 5 = Denhardt’s, 250 mgrml yeast tRNA, 500 mgrml salmon sperm DNA. containing 0.5 ngrml heat-denaturated DIG-labeled cRNA probe. Slides were then processed as previously described w18x. DIG was detected using alkaline phosphatase-conjugated anti-DIG antibody, 4-nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indoyl-phosphate. Sections were then processed for Phox2a immunohistochemistry, as described above.
3. Results 3.1. Apoptosis in the LC during postnatal deÕelopment Seven postnatal ages ŽP13, P14, P16, P19, P21, P23, and P30. were selected to monitor, in the rostral pole of the LC proper, the presence of cell death, by labeling DNA breaks using the ISEL method. Among all sections analyzed, three cells only exhibiting DNA breaks were observed in the LC, one at P14 ŽFig. 2A. and one at P16 in the ‘‘core area of the LC’’, and one at P16 in the ‘‘extracore area’’ of the LC. The ponctuata appearance of these dying cells and the high condensation of the labeled material ŽFig. 2B. indicate that cell death likely occurred via apoptosis. 3.2. TH-mRNA concentration in subsets of neurons in the rostral pole of LC proper at P14 The detection of TH-mRNA by in situ hybridization at P23 ŽFig. 3B. confirmed the disappearance of TH gene
278
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
tissue concentration increased 66 " 23% Ž p - 0.05. and 110 " 28% Ž p - 0.01. above controls, 1 and 3 days after treatment, respectively. Three days after treatment, at the rostral pole of the LC proper, the number of TH-mRNAcontaining perikarya increased, causing an ‘‘enlargement’’ of the nucleus and the emergence of an ‘‘extra-core area’’ of the LC ŽFig. 3B,C.. The density of specific silver grains in the LC perikarya, which was used as an index of TH-mRNA perikaryal concentration, was normally distributed across the populations of cells selected in the ‘‘core area of the LC’’, in control rats Žnot shown. and RU 24722-treated rats Ž n s 52, Pearson x 2 test of normality, Fig. 3F.. In the ‘‘extra-core area’’ of the LC, which was revealed in RU 24722-treated rats, the silver grain density was not normally distributed ŽFig. 3G.. Fig. 2. In situ detection of apoptotic neurons in the rostral pole of the LC at P14. ŽA. On this Nissl-stained section, a single apoptotic neuron Žpointed by the arrow. was detected in the LC Žoutlined area., and was located in the ‘‘core area’’ of the nucleus. ŽB. Observation of the apoptotic neuron pointed in panel A at a higher magnification, demonstrating dense condensation of the DNA. Bars: 100 mm ŽA.; 20 mm ŽB..
expression in the cells which are localized, at P14, in the ventro-lateral tip of the LC, the so-called ‘‘extra-core area’’ ŽFig. 3A.. The density of silver grains corresponding to non-specific labeling was determined in areas ventral to the LC and ranged between 0.8% and 1.4% of the silver grain density determined in the perikarya of LC neurons. The density of specific silver grains in the perikarya selected randomly for evaluation was normally distributed across the population of cells selected in the ‘‘core area of the LC’’ Ž n s 52, Pearson x 2 test of normality. ŽFig. 3D., but this density was not normally distributed in the ‘‘extra-core area’’ of the LC ŽFig. 3E.. 3.3. TH gene expression in response to RU 24722 treatment at P14 and P23 We investigated in P14 RU 24722-treated rats, whether additional TH-expressing cells could be detected at the vicinity of LC cells which expressed TH gene in basal conditions. The topology of the LC delineated by TH-IP perikarya was reconstructed at P14 in control rats and in rats which had received a single dose of RU 24722 at P11. We found that no alteration had occurred in RU 24722treated rats, both in the surface area delineated by TH-IP perikarya and the topology of the LC as compared to controls Žnot shown.. Pharmacologically induced alterations in TH-mRNA tissue concentration in the LC were monitored at P23 by in situ hybridization and film autoradiography. TH-mRNA tissue concentration was measured 1 and 3 days after a single administration of RU 24722, in four anatomical planes selected at the caudo-rostral junction of the LC and separated by 160 mm. In this 640-mm interval, TH-mRNA
3.4. Phox2a expression in the LC of control and RU 24722-treated adult rats The anatomical distribution of Phox2a-IP cells was heterogeneous along the caudo-rostral axis of the LC; the majority Ž; 80%. of stained cells was contained in the half-caudal subdivision of the structure Žanatomical planes 1 to 4, Fig. 4A.. The mean diameter d of labeled nuclei was 10.7 mm Ž n s 71; 8.2 - d - 13.1 mm.. The number of Phox2a-IP cells, estimated in the entire structure using the Abercrombie formula, was 2588 " 164 in basal conditions. In situ dual labeling of TH-mRNA and Phox2a protein revealed that all TH-IP neurons expressed Phox2a and that numerous Phox2a-IP cells did not contain detectable level of TH-mRNA; 47% of total Phox2a-IP cells did not contain TH at anatomical plane 4 ŽFig. 4B.. RU 24722 treatment did not change the caudo-rostral distribution of Phox2a-IP cells in the LC ŽFig. 4A., nor the total number of Phox2a-IP cells contained in the entire structure Ž2453 " 122..
4. Discussion 4.1. Apoptotic death is not preÕalent in the deÕeloping LC Excess numbers of neurons are produced during the ontogenesis within the central nervous system, and are eliminated, often having served important but transient functions w15x. The noradrenergic innervation of the brain is progressive during pre- and postnatal development, and the noradrenaline released by LC neurons may be involved in the critical periods of brain plasticity and regulate different brain development steps w2,10,13,16x. We previously reported that the number of TH-expressing neurons at P14 was 33% greater than that at P42 w4x. If the function served by these additional TH-expressing neurons was transient during development, the loss of TH phenotype in
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
279
Fig. 3. Analysis of TH-mRNA presence at the rostral pole of LC proper. ŽA–C. Polarized light views of emulsion-coated coronal sections processed for TH-mRNA in situ hybridization using a 35 S-labelled oligonucleotidic probe, at P14 ŽA. and at P23 ŽB, C., in control ŽA, B. and RU 24722-treated rats ŽC.. Dashed lines indicate the position of the ventral limit of the 4th ventricle. Bar: 300 mm ŽA, B, C.. ŽD–G. Frequency histograms illustrating the distribution of an index of TH-mRNA perikaryal concentration from individual perikarya selected within the ‘‘core area’’ ŽD, F. and the ‘‘extra-core area’’ ŽE, G., in P14 control rats ŽD, E. and P23 RU 24722-treated rats ŽF, G.. At P23, the ‘‘core area’’ only was detected in control rats and its topology has been quantitatively reconstructed. The superimposition of this reconstructed ‘‘core area’’ over each section analyzed at P14, and at P23 in RU 24722-treated rats, made it possible to isolate the ‘‘core area’’ and the ‘‘extra-core area’’ of the LC. Index of TH-mRNA perikaryal concentration was expressed as the density of silver grains ŽAU. over selected neuron perikarya of four animals. Total number n of neurons considered is indicated. The index of TH-mRNA perikaryal concentration was ‘‘normally’’ distributed across the population of neurons selected in the ‘‘core area’’ of the LC only.
these neurons beyond P14 may reflect either TH gene repression or massive cell death. Three apoptotic cells only were detected at the rostral pole of the LC proper throughout the time period of postnatal development examined ŽP13 to P30.. Since numerous apoptotic cells were observed at P13 in the substantia nigra Žnot shown., as previously reported w9x, the methodology we used can be considered as sensitive and reliable enough to detect apoptotic cells in brain sections. No cells were double-labeled
when dual ISEL and immunohistochemical detection of TH was performed. This may indicate that dying cells did not express TH or that the epitopes of TH protein recognized by the different antibodies raised against TH that we used were already altered when immunohistochemistry was performed. Anyway, the present data strongly suggest that apoptotic cell death was not a massive phenomenon within the rostral pole of the LC proper throughout the postnatal development of the rat. Therefore, loss of TH
280
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
phenotype beyond P14 may reflect TH gene repression rather than cell death. 4.2. TH-mRNA concentration reÕeals transient TH phenotype in subsets of LC neurons We investigated whether it was possible to predict transient expression of TH phenotype in the ‘‘extra-core area’’ of the LC in development or in adults treated with RU 24722. Therefore, we compared some modalities of TH gene expression in the ‘‘extra-core area’’ with that of the ‘‘core area’’, which contains LC neurons expressing TH permanently in basal conditions throughout development. TH-mRNA was chosen as a marker to examine, since, in contrast with TH protein, its subcellular distribution is restricted to the perikarya of rat LC neurons w3x, making it possible to measure its concentration in individual neurons. We observed that the concentration of THmRNA exhibited a profile of ‘‘normal’’ distribution in the population of cells localized in the ‘‘core area’’ of the LC, in any conditions examined, but not in the ‘‘extra-core area’’ of the LC both at P14 in control LC and at P23 in the LC of rats treated with RU 24722 at P20. The observation of a bi-modal profile in the ‘‘extra-core area’’ probably means that neurons were progressively recruited for transcription of the TH gene, as supported by the observation that there is no synchronous transcription activity of TH gene among the population of LC neurons w19x. In the ‘‘extra-core area’’ of the LC, cells with low concentration of TH-mRNA may have just started to express TH or may have already started the process of TH gene repression at the time when measurements were performed. 4.3. Expression of homeoprotein Phox2a in ‘‘restingnoradrenergic’’ adult LC neurons
Fig. 4. Expression of Phox2a homeoprotein in the LC of adult rats in basal conditions and 3 days after a single treatment with RU 24722. ŽA. Number of Phox2a-IP cells counted in 20-mm-thick sections, selected every 200 mm along the caudo-rostral extent of the LC. Results are expressed as the means"S.E.M. Ž ns 3., for both control Žwhite bars. and RU 24722-treated Ždark bars. rats. ŽB. Color photomicrograph of a section through the rostral LC of control rat, from material processed for dual in situ hybridization of TH-mRNA and immunohistochemical detection of Phox2a. TH-mRNAq cells are apparent by the presence of blue reaction product in their cytoplasm and Phox2aq cells by that of brown reaction product in their nucleus. At this anatomical plane Žnumber 4, as in A., Phox2a-IP cells without detectable expression of TH-mRNA are pointed by arrowheads. 4v: 4th ventricle. Bar: 50 mm.
Homeodomain transcription factor Phox2a is expressed in domains within the nervous system corresponding to noradrenergic cell populations w18x, and appears to be a determinant of the noradrenergic phenotype w7,14x. We found in the present study that numerous cells in the LC expressed Phox2a, but not TH in basal conditions, suggesting the existence of a ‘‘reservoir’’ of cells within the LC with the intrinsic capacity to develop noradrenergic phenotype in response to pertinent stimuli. Indeed, we previously reported ; 1700 TH-positive neurons in the adult rat LC w4,5x, contrasting with the ; 2600 Phox2a-positive cells counted in the present study. The observation that the number of Phox2a-expressing cells did not increase following RU 24722 treatment, as does the number of TH-expressing cells w5x, indicates that TH gene expression may have been triggered in response to the treatment in neurons that already expressed Phox2a in basal conditions. Interestingly, the number of Phox2a positive cells was greater than the number Ž; 2200 w5x. of TH-positive cells counted in the LC following RU 24722. Thus, subsets of Phox2a-ex-
L. Bezin et al.r Molecular Brain Research 76 (2000) 275–281
pressing cells still do not express TH in response to RU 24722 treatment, and may correspond to the cells that lost TH phenotype between P21 and P42 ŽFig. 1B.. Thus, Phox2a is not sufficient to activate TH gene expression in singular subsets of LC neurons in basal conditions. However, Phox2a has been demonstrated to induce expression of TH in collaboration with cAMP w11x. Further studies will be needed to evaluate the involvement of cAMP in the plasticity of TH phenotype in LC neurons expressing Phox2a. 5. Conclusion In attempting to determine whether the cells in which TH expression was repressed between P14 and P21 could correspond to those which exhibited TH phenotype in response to RU 24722 treatment in adults, we verified that no massive cell death occurred in the LC during the postnatal period. We also observed that both cell populations exhibited the same altered steady-state concentration of TH-mRNA as compared to cells which permanently expressed TH. Finally, we have demonstrated the presence of TH-negative neurons expressing the homeodomain transcription factor Phox2a, specific for noradrenergic neurons, providing further evidence that ‘‘restingnoradrenergic’’ neurons exist in the adult rat LC under basal conditions. The existence of these numerous cells, localized in the LC where efferents have been shown to be widely spread across the central nervous system, provides interesting prospectives for gain of noradrenergic function when classical noradrenergic LC neurons are impaired. Acknowledgements This work was supported by grants from the Centre National de la Recherche Scientifique ŽERS 2022. and the Universite´ Claude Bernard Lyon I. The authors are very grateful to Drs. C. Goridis and J.-Fr. Brunet for their critical reading of the manuscript. References w1x M. Abercrombie, Estimation of nuclear population from microtome sections, Anat. Rec. 94 Ž1946. 239–247. w2x M.F. Bear, W. Singer, Modulation of visual cortical plasticity by acetylcholine and noradrenaline, Nature 320 Ž1986. 172–176.
281
w3x L. Bezin, J.J. Diaz, D. Marcel, M. Le Cavorsin, J.J. Madjar, J.F. Pujol, D. Weissmann, Controlled targeting of tyrosine hydroxylase protein toward processes of locus coeruleus neurons during postnatal development, Mol. Brain Res. 50 Ž1997. 23–32. w4x L. Bezin, D. Marcel, L.I. Debure, N. Ginovart, C. Rousset, J.F. Pujol, D. Weissmann, Postnatal development of the tyrosine hydroxylase-containing cell population within the rat locus coeruleus: topological organization and phenotypic plasticity, J. Neurosci. 14 Ž1994. 7486–7501. w5x L.I. Debure, L. Bezin, N. Ginovart, C. Rousset, J.F. Pujol, D. Weissmann, RU24722 induces spatially organized phenotypic plasticity in the locus coeruleus, NeuroReport 5 Ž1994. 1793–1796. w6x L.I. Debure, E. Moyse, M. Fevre-Montange, H. Hardin, M.F. Belin, C. Rousset, J.F. Pujol, D. Weissmann, Somatotopic organization of tyrosine hydroxylase in the rat locus coeruleus: long-term effect of RU24722, Brain Res. 581 Ž1992. 19–32. w7x C. Goridis, J.F. Brunet, Transcriptional control of neurotransmitter phenotype, Curr. Opin. Neurobiol. 9 Ž1999. 47–53. w8x B. Grima, A. Lamouroux, F. Blanot, N. Faucon-Biguet, J. Mallet, Complete coding sequence of rat tyrosine hydroxylase mRNA, Proc. Natl. Acad. Sci. U. S. A. 82 Ž1985. 617–621. w9x E. Janec, R.E. Burke, Naturally occurring cell death during postnatal development of the substantia nigra pars compacta of rat, Mol. Cell. Neurosci. 4 Ž1993. 30–35. w10x T. Kasamatsu, J.D. Pettigrew, Preservation of binocularity after monocular deprivation in the striate cortex of kittens treated with 6-hydroxydopamine, J. Comp. Neurol. 185 Ž1979. 139–162. w11x L. Lo, X. Morin, J.F. Brunet, D.J. Anderson, Specification of neurotransmitter identity by Phox2 proteins in neural crest stem cells, Neuron 22 Ž1999. 693–705. w12x S.E. Loughlin, S.L. Foote, F.E. Bloom, Efferent projections of nucleus locus coeruleus: topographic organization of cells of origin demonstrated by three-dimensional reconstruction, Neuroscience 18 Ž1986. 291–306. w13x T. Maeda, M. Tohyama, N. Shimizu, Modification of postnatal development of neocortex in rat brain with experimental deprivation of locus coeruleus, Brain Res. 70 Ž1974. 515–520. w14x X. Morin, H. Cremer, M.R. Hirsch, R.P. Kapur, C. Goridis, J.F. Brunet, Defects in sensory and autonomic ganglia and absence of locus coeruleus in mice deficient for the homeobox gene Phox2a, Neuron 18 Ž1997. 411–423. w15x R.W. Oppenheim, Cell death during development of the nervous system, Annu. Rev. Neurosci. 14 Ž1991. 453–501. w16x J.G. Parnavelas, M.E. Blue, The role of the noradrenergic system on the formation of synapses in the visual cortex of the rat, Dev. Brain Res. 3 Ž1982. 140–144. w17x L.W. Swanson, The locus coeruleus: a cytoarchitectonic, golgi and immunocytochemical study in the albino rat, Brain Res. 110 Ž1976. 39–56. w18x M.C. Tiveron, M.R. Hirsch, J.F. Brunet, The expression pattern of the transcription factor Phox2 delineates synaptic pathways of the autonomic nervous system, J. Neurosci. 16 Ž1996. 7649–7660. w19x A. Trembleau, F.E. Bloom, Subcellular localization of tyrosine hydroxylase ŽTH. gene transcripts: new insights into the pattern of TH gene expression in the locus coeruleus under pharmacological stimulation, Biol. Cell 90 Ž1998. 39–51.